A real-time acquisition and detection method and system for frequency hopping signals

Abstract

The application relates to a real-time acquisition detection method and system for frequency hopping signals, which comprises the following steps: obtaining frequency spectrum diagram data in a certain time period according to received and buffered STFT data, analyzing the frequency spectrum diagram data to obtain all suspected signals; performing graph model establishment on all suspected signals through a frequency hopping sorting algorithm, and judging whether a frequency hopping signal exists according to the graph model; if no frequency hopping signal is detected, cleaning the received and buffered IQ data, and continuing to detect the frequency hopping signal in the buffered data; if a frequency hopping signal is detected, storing the IQ data in the buffered queue and continuing to detect the frequency hopping signal in the buffered data. The application can quickly judge whether a frequency hopping signal exists, the frequency spectrum diagram is processed in a matrix mode, sparse matrix features are extracted, the feature matrix is efficiently compressed and processed, and parameter information of the signal is obtained.

Description

Technical Field

[0001] This invention relates to the field of signal detection technology, and in particular to a real-time acquisition and detection method and system for frequency hopping signals. Background Technology

[0002] Frequency-hopping communication has been under research both domestically and internationally since the 1960s, focusing on new spread spectrum communication systems, theories, and technologies. In recent years, its application in civilian communication systems, particularly cellular mobile phone systems, has gained increasing attention. Frequency-hopping technology has also been widely used in wireless local area networks (WLANs), such as Bluetooth, which is built upon it. With the continuous development of communication technology, frequency-hopping technology has been increasingly applied, further exposing the inefficiency of manual data collection. Therefore, how to automate the collection of frequency-hopping signal data has become a growing concern for engineers.

[0003] In recent years, computer vision technology has matured and has been applied to more and more industries with excellent results. Combining computer vision-related technologies with radio signal detection has become the choice of more and more signal researchers. Based on traditional data acquisition methods, how to achieve "what you see is what you get" and timely acquisition of frequency-hopping signals, replacing manual acquisition processes with automation, has become an increasingly urgent issue for technicians. At the same time, automated acquisition needs to have good robustness under various noise and interference conditions and be applicable to various frequency-hopping signals, which places increasingly higher demands on the automation process.

[0004] It should be noted that the information disclosed in the background section above is only used to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art and provide a real-time acquisition and detection method and system for frequency hopping signals, which solves the problems existing in the current manual acquisition.

[0006] The objective of this invention is achieved through the following technical solution: a real-time acquisition and detection method for frequency hopping signals, the acquisition and detection method comprising:

[0007] The receiver with the antenna outputs STFT data and IQ data in real time, buffers the data, obtains the spectrum data for a certain period of time based on the received and buffered STFT data, and analyzes the spectrum data to obtain all suspected signals.

[0008] A graphical model is built for all suspected signals using a frequency hopping sorting algorithm, and the presence of frequency hopping signals is determined based on the graphical model.

[0009] If no frequency hopping signal is detected, the received and buffered IQ data is cleared, and frequency hopping signal detection processing continues for the buffered data. If a frequency hopping signal is detected, the IQ data in the buffer queue is stored on disk, and frequency hopping signal detection processing continues for the buffered data.

[0010] The analysis of the spectrogram data yielded all suspected signals, including:

[0011] The back noise amplitude is obtained by density clustering of the spectrum, and the signal filtering threshold is set according to the back noise amplitude.

[0012] The spectrum is filtered based on the signal threshold to obtain a 0 / 1 feature matrix.

[0013] The start time, end time, start frequency, and end frequency of the signal are obtained based on time-based signal continuity analysis.

[0014] The method of obtaining the back noise amplitude by density clustering of the spectrum includes:

[0015] For the spectrum Figure 2 Statistical analysis of the amplitude is performed to transform the problem of solving the background amplitude problem into a one-dimensional density clustering problem. The location where the amplitude density is most concentrated is obtained by using machine learning density clustering methods, and this location is the background amplitude noise.

[0016] The step of establishing a graphical model for all suspected signals using a frequency hopping sorting algorithm, and then using the graphical model to search for and determine whether a frequency hopping signal exists, specifically includes:

[0017] Extract a signal from the input signal set, iterate through the other signals, and if the end time of the other signals is the same as the start time of the extracted signal or the start time of the other signals is the same as the end time of the extracted signal, then add it to the graph sequence of the extracted signals. Then perform the same operation on the newly added signals to construct the graph model of the initial extracted signals.

[0018] The longest path is solved for the graph model to obtain a sequence of suspected frequency hopping signals. The length of the sequence is judged. If the number of frequency hopping is greater than or equal to the threshold, it is added to the output set of frequency hopping sequences.

[0019] Remove the signals from the signal set from the graphical model and repeat the above operation to obtain the set of all frequency hopping sequences.

[0020] A real-time acquisition and detection system for frequency hopping signals includes a data caching module, a density clustering analysis module, a frequency hopping sorting module, and a result processing module.

[0021] The data caching module: The receiver outputs STFT data and IQ data in real time through the antenna, caches the data, and obtains the spectrum data within a certain time period based on the received and cached STFT data;

[0022] The density clustering analysis module is used to analyze the spectrogram data using density clustering to obtain all suspected signals.

[0023] The frequency hopping sorting module is used to establish a graphical model for all suspected signals using a frequency hopping sorting algorithm, and to search for and determine whether a frequency hopping signal exists based on the graphical model.

[0024] The result processing module is used to: if no frequency hopping signal is detected, clean up the received and cached IQ data and continue to process the cached data for frequency hopping signal detection; if a frequency hopping signal is detected, store the IQ data in the cache queue to disk and continue to process the cached data for frequency hopping signal detection.

[0025] The density clustering analysis module includes a density clustering unit, a threshold filtering unit, and a continuity analysis unit;

[0026] The density clustering unit is used for spectrum analysis. Figure 2 Statistical analysis of the amplitude is performed to transform the problem of solving the background amplitude problem into a one-dimensional density clustering problem. The location where the amplitude density is most concentrated is obtained by machine learning density clustering, and this location is the background amplitude noise.

[0027] The threshold filtering unit is used to set a signal filtering threshold according to the back noise amplitude, and to filter the spectrum according to the signal threshold. When the spectrum amplitude is greater than the signal filtering threshold, it is set to 1, otherwise it is set to 0, thus obtaining a 0 / 1 feature matrix.

[0028] The continuity analysis unit is used to obtain the start time, end time, start frequency, and end frequency information of the signal based on time-based signal continuity analysis.

[0029] The frequency hopping sorting module includes a graph model construction unit, a graph model analysis and processing unit, and a loop unit;

[0030] The graph model construction unit is used to extract a signal from the input signal set, traverse other signals, and add the extracted signal to the graph sequence if the end time of other signals is the same as the start time of the extracted signal or the start time of other signals is the same as the end time of the extracted signal. Then, the corresponding operation is performed on the newly added signal to construct the graph model of the initial extracted signal.

[0031] The graph model analysis and processing unit is used to solve the longest path problem of the graph model to obtain a sequence of suspected frequency hopping signals, and to judge the length of the sequence. If the number of frequency hopping is greater than or equal to the threshold, it is put into the output set of frequency hopping sequences.

[0032] The loop unit is used to remove signals from the signal set in the graphical model and repeat the operations of the graphical model construction unit and the graphical model analysis and processing unit to obtain all frequency hopping sequence sets.

[0033] The present invention has the following advantages:

[0034] 1. It has strong generalization ability for frequency hopping signals. It can achieve good results for conventional frequency hopping signals; current technologies rely heavily on the characteristic that conventional frequency hopping signals have equal hopping rates, while this invention has a certain effect on frequency hopping signals under various hopping rate conditions; at the same time, this method can adapt to frequency hopping signals under different bandwidths, and has strong generalization ability.

[0035] 2. It has strong anti-noise and anti-interference capabilities. It detects signals through spectrum analysis and can adapt to various interference and noise conditions, achieving what you see is what you get. At the same time, according to the data disk acquisition mechanism, when a frequency hopping signal is detected, it will be disk-acquired for a certain period of time before and after that period. Even if there are missing hopping signals, the signal can still be acquired relatively completely.

[0036] 3. It can quickly determine whether a frequency hopping signal exists. It processes the spectrum using a matrix approach and extracts sparse matrix features. By efficiently compressing and processing the feature matrix, it obtains the signal's parameter information. At the same time, it uses a graphical model to analyze the frequency hopping sequence of all suspected signals, and finally determines whether the frequency hopping signal exists. Attached Figure Description

[0037] Figure 1 This is a schematic diagram of the overall process of the frequency hopping signal acquisition method of the present invention.

[0038] Figure 2 This is a schematic diagram of the suspected signal detection process of the present invention.

[0039] Figure 3 This is a schematic diagram of the overall process of the frequency hopping sorting algorithm of the present invention. Detailed Implementation

[0040] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the detailed description of the embodiments of this application provided below with reference to the accompanying drawings is not intended to limit the scope of protection of the claimed application, but merely represents selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application. The present invention will be further described below with reference to the accompanying drawings.

[0041] This invention specifically relates to a method for real-time active detection and automatic data acquisition of frequency-hopping signals. This method can analyze and judge real-time signal data, and upon detecting a frequency-hopping signal in the real-time data, can store the signal's IQ data in real time. It includes the following:

[0042] Step 1: The receiver connected to the antenna outputs STFT (Short Time Fourier Transform) data and IQ data in real time and buffers the data.

[0043] Step 2: After the STFT data is cached, the spectrum data within a certain time period can be obtained. The background noise amplitude is obtained by density clustering, and the signal filtering threshold is set according to the background noise amplitude.

[0044] Step 3: Filter the spectrum according to the signal threshold to obtain the 0 / 1 feature matrix;

[0045] Step 4: Obtain the suspected signals at each frequency point based on time-based signal continuity analysis.

[0046] Step 5: Then, based on the frequency-based signal continuity analysis, obtain the signal parameter information, including the signal's start time, end time, start frequency, and end frequency.

[0047] Step 6: First, establish a graphical model for all suspected signals using a frequency hopping sorting algorithm, then search for frequency hopping signals based on the graphical model, and finally output whether a frequency hopping signal exists in the spectrum graph.

[0048] Step 7: If no frequency hopping signal is detected, the IQ cache data is cleared, and frequency hopping signal detection processing continues for the cache data; if a frequency hopping signal is detected, the IQ data in the cache queue is stored on disk, and frequency hopping signal detection processing continues for the cache data.

[0049] Furthermore, such as Figure 1As shown, the receiver connected to the antenna first outputs STFT data and IQ data in real time and buffers them. After the STFT data is buffered, a spectrum data for a certain period of time can be obtained. The spectrum data is analyzed to obtain all suspected signals. A frequency hopping signal sorting algorithm is used to determine whether a frequency hopping signal exists. If no frequency hopping signal is detected, the IQ buffer data is cleared and the frequency hopping signal detection processing continues for the buffer data. If a frequency hopping signal is detected, the IQ data in the buffer queue is stored on disk and the frequency hopping signal detection processing continues for the buffer data.

[0050] Furthermore, such as Figure 2 As shown, the background noise amplitude is first obtained from the spectrum using density clustering. This is achieved by... Figure 2 Statistical analysis of amplitude is performed, transforming the problem of solving the background amplitude problem into a one-dimensional density clustering problem. Using the traditional machine learning density clustering method DBSCAN, the location with the highest amplitude density can be found; this location represents the background amplitude noise. A signal filtering threshold is set based on the background noise amplitude. The spectrogram is filtered based on the signal threshold; if the spectrogram amplitude is greater than the signal filtering threshold, it is set to 1; otherwise, it is set to 0, resulting in a 0 / 1 feature matrix. Suspected signals at each frequency point are obtained based on time-based signal continuity analysis. Furthermore, signal parameter information, including the signal's start time, end time, start frequency, and end frequency, is obtained based on frequency-based signal continuity analysis. For example, when analyzing the time-axis sequence of the spectrogram using continuity analysis, if multiple consecutive 1s appear in the feature graph, a signal is considered to exist in that region, and the start and end times of the signal can be determined.

[0051] Furthermore, such as Figure 3 As shown, a graphical model is first built for all suspected signals, and frequency-hopping signals are searched based on the graphical model. The final output is whether a frequency-hopping signal exists in the spectrum. The algorithm's input is a set of signals and a threshold for the number of consecutive frequency-hopping signals. The signal set includes the start and end times and start and end frequencies of the signals. The algorithm's output is a set of frequency-hopping sequences. The algorithm first extracts a signal from the signal set and iterates through the other signals. If the end time of another signal is the same as the start time of the selected signal, or if the start time of another signal is the same as the end time of the selected signal, it is added to the graphical sequence of the extracted signal. This process is repeated for newly added signals to construct the initial graphical model of the selected signals. The longest path is then solved on this graphical model to obtain the sequence of suspected frequency-hopping signals. The sequence length is judged; if the number of frequency hopping signals is greater than or equal to the threshold, it is added to the output set of frequency-hopping sequences. Signals in the graphical model are then removed from the signal set. This process is repeated to obtain the complete set of frequency-hopping sequences.

[0052] The above description is merely a preferred embodiment of the present invention. It should be understood that the present invention is not limited to the forms disclosed herein and should not be construed as excluding other embodiments. It can be used in various other combinations, modifications, and improvements, and can be altered within the scope of the concept described herein through the above teachings or related technologies or knowledge. Modifications and variations made by those skilled in the art that do not depart from the spirit and scope of the present invention should be within the protection scope of the appended claims.

Claims

1. A real-time acquisition and detection method for frequency hopping signals, characterized in that: The acquisition and detection method includes: The receiver outputs STFT and IQ data in real time through the antenna, buffers the data, obtains the spectrum data for a certain period of time based on the received and buffered STFT data, and analyzes the spectrum data to obtain all suspected signals; among them, suspected signals are burst signals of suspected frequency hopping signals. A graphical model is built for all suspected signals using a frequency hopping sorting algorithm, and the presence of frequency hopping signals is determined based on the graphical model. If no frequency hopping signal is detected, the received and buffered IQ data is cleared, and frequency hopping signal detection processing continues for the buffered data. If a frequency hopping signal is detected, the IQ data in the buffer queue is stored on disk, and frequency hopping signal detection processing continues for the buffered data. The step of establishing a graphical model for all suspected signals using a frequency hopping sorting algorithm, and then using the graphical model to search for and determine whether a frequency hopping signal exists, specifically includes: A1. Extract a signal from the input signal set, iterate through the other signals, and if the end time of the other signals is the same as the start time of the extracted signal or the start time of the other signals is the same as the end time of the extracted signal, then add it to the graph sequence of the extracted signals. Then perform the same operation on the newly added signals to construct the graph model of the initial extracted signals. A2. Solve the longest path problem for the graph model to obtain the sequence of suspected frequency hopping signals. Determine the length of the sequence. If the number of frequency hopping is greater than or equal to the threshold, add it to the output set of frequency hopping sequences. A3. Remove the signals from the signal set in the graphical model, and repeat the operations of A1 and A2 to obtain the set of all frequency hopping sequences.

2. The real-time acquisition and detection method for frequency hopping signals according to claim 1, characterized in that: The analysis of the spectrogram data yielded all suspected signals, including: The back noise amplitude is obtained by density clustering of the spectrum, and the signal filtering threshold is set according to the back noise amplitude. The spectrum is filtered based on the signal threshold to obtain a 0 / 1 feature matrix. The start time, end time, start frequency, and end frequency of the signal are obtained based on time-based signal continuity analysis.

3. The real-time acquisition and detection method for frequency hopping signals according to claim 2, characterized in that: The method of obtaining the back noise amplitude by density clustering of the spectrum includes: By performing statistical analysis on the two-dimensional amplitude of the spectrogram, the problem of solving the background amplitude is transformed into a one-dimensional density clustering problem. The location where the amplitude density is most concentrated is obtained by using machine learning density clustering, and this location is the background amplitude noise.

4. A real-time acquisition and detection system for frequency hopping signals, characterized in that: It includes a data caching module, a density clustering analysis module, a frequency hopping sorting module, and a result processing module; The data caching module: The receiver outputs STFT data and IQ data in real time through the antenna, caches the data, and obtains the spectrum data within a certain time period based on the received and cached STFT data; The density clustering analysis module is used to analyze the spectrogram data using density clustering to obtain all suspected signals. The frequency hopping sorting module is used to establish a graphical model for all suspected signals using a frequency hopping sorting algorithm, and to search for and determine whether a frequency hopping signal exists based on the graphical model. The result processing module is used to: if no frequency hopping signal is detected, clean up the received and buffered IQ data and continue to process the buffered data for frequency hopping signal detection; if a frequency hopping signal is detected, store the IQ data in the buffer queue onto disk and continue to process the buffered data for frequency hopping signal detection. The frequency hopping sorting module includes a graph model construction unit, a graph model analysis and processing unit, and a loop unit; The graph model construction unit is used to extract a signal from the input signal set, traverse other signals, and add the extracted signal to the graph sequence if the end time of other signals is the same as the start time of the extracted signal or the start time of other signals is the same as the end time of the extracted signal. Then, the corresponding operation is performed on the newly added signal to construct the graph model of the initial extracted signal. The graph model analysis and processing unit is used to solve the longest path problem of the graph model to obtain a sequence of suspected frequency hopping signals, and to judge the length of the sequence. If the number of frequency hopping is greater than or equal to the threshold, it is put into the output set of frequency hopping sequences. The loop unit is used to remove signals from the signal set in the graphical model and repeat the operations of the graphical model construction unit and the graphical model analysis and processing unit to obtain all frequency hopping sequence sets.

5. The real-time acquisition and detection system for frequency hopping signals according to claim 4, characterized in that: The density clustering analysis module includes a density clustering unit, a threshold filtering unit, and a continuity analysis unit; The density clustering unit is used to perform statistical analysis on the two-dimensional amplitude of the spectrum, transforming the problem of solving the background amplitude into a one-dimensional density clustering problem. The location where the amplitude density is most concentrated is obtained through machine learning density clustering, and this location is the background amplitude noise. The threshold filtering unit is used to set a signal filtering threshold according to the back noise amplitude, and to filter the spectrum according to the signal threshold. When the spectrum amplitude is greater than the signal filtering threshold, it is set to 1, otherwise it is set to 0, thus obtaining a 0 / 1 feature matrix. The continuity analysis unit is used to obtain the start time, end time, start frequency, and end frequency information of the signal based on time-based signal continuity analysis.

Images